Isolation and Characterization of Phosphate Solubilizing Bacteria from Paddy Field Soils in Japan

Phosphorus (P) is abundant in soil and is essential for plant growth and development; however, it is easily rendered insoluble in complexes of different types of phosphates, which may lead to P deficiency. Therefore, increases in the amount of P released from phosphate minerals using microbial inoculants is an important aspect of agriculture. The present study used inorganic phosphate solubilizing bacteria (iPSB) in paddy field soils to develop microbial inoculants. Soils planted with rice were collected from different regions of Japan. Soil P was sequentially fractionated using the Hedley method. iPSB were isolated using selective media supplemented with tricalcium phosphate (Ca-P), aluminum phosphate (Al-P), or iron phosphate (Fe-P). Representative isolates were selected based on the P solubilization index and soil sampling site. Identification was performed using 16S rRNA and rpoB gene sequencing. Effectiveness was screened based on rice cultivar Koshihikari growth supplemented with Ca-P, Al-P, or Fe-P as the sole P source. Despite the relatively homogenous soil pH of paddy field sources, three sets of iPSB were isolated, suggesting the influence of fertilizer management and soil types. Most isolates were categorized as β-Proteobacteria (43%). To the best of our knowledge, this is the first study to describe the genera Pleomorphomonas, Rhodanobacter, and Trinickia as iPSB. Acidovorax sp. JC5, Pseudomonas sp. JC11, Burkholderia sp. JA6 and JA10, Sphingomonas sp. JA11, Mycolicibacterium sp. JF5, and Variovorax sp. JF6 promoted plant growth in rice supplemented with an insoluble P source. The iPSBs obtained may be developed as microbial inoculants for various soil types with different P fixation capacities.

Phosphorus (P) is essential for plant growth and development. Although there is an abundance of P in the soil, only 0.1% of total P (TP) is available for plant uptake (Zou et al., 1992). P is easily rendered insoluble in complexes of different types of phosphates. Stable minerals, such as variscite (AlPO 4 .2H 2 O) and strengite (FePO 4 .2H 2 O), are commonly present in acidic agricultural soils (Richardson, 2001). In addition, soluble orthophosphate is rarely detected in alkaline soils, which contain abundant amounts of calcium (Goldstein et al., 1999). The insoluble nature of P leads to P deficiency and limits crop productivity. Furthermore, there has been a depletion of P reserves (Cordell et al., 2009) because it is a non-renewable source and there is an increasing demand for the bulk production of fertilizers (Dorozhkin, 2011). The production and utilization processes associated with P fertilizers are environmentally undesirable and costly because of the high prices driven by increasing demand, which may place a burden on farmers (Vassilev et al., 2006). Furthermore, the excessive usage of chemical P fertilizers poses a threat to the environment, specifically water pollution (Bennett et al., 2001). An increased dependence on chemical fertilizers and pesticides poses environmental issues, such as air and groundwater pollution, due to the eutrophication of water bodies (Youssef and Eissa, 2014). Therefore, phosphate fertilization is a major agricultural research matter.
An efficient, low-cost, and environmentally friendly approach to soil P fertilization needs to be developed. Specific species of microorganisms with unique properties are often used as alternatives to reduce the use of chemical fertilizers (Deepali and Gangwar, 2010). Phosphate solubilizing microorganisms (PSM) convert insoluble P compounds into soluble forms that are suitable for plant uptake by metabolizing root-borne C compounds, mainly sugars and organic acids (Rodrıǵuez and Fraga, 1999;Whitelaw, 2000;Arcand and Schneider, 2006). Some species of Pseudomonas, Enterobacter, Serratia, Pantoea, Aspergillus, and Penicillium actively solubilize phosphate (Whitelaw, 2000;Son et al., 2006;Buch et al., 2008;Gulati et al., 2008;Sulbarán et al., 2009). These bacteria and fungi grow in media with insoluble phosphate compounds as the sole source of phosphate. PSM not only assimilate P, they also release a large portion of soluble P in excess of their own requirements. The solubilization of mineral phosphates is an important plant growth-promoting (PGP) ability and PSM are now being used as inoculants for crop production.
In Japan, the application of chemical P fertilizers decreased between 1985 and 2005 (Mishima et al., 2010). For example, the application of chemical P fertilizers in paddy fields decreased from 117.0 kg P ha -1 in 1985 to 83.6 kg P ha -1 in 2005. Although there was a decline in the utilization of P fertilizers, a reduction in P uptake by paddy rice was not observed. This finding indicated that the application of chemical P fertilizers may be reduced without adverse effects on paddy rice yield and was attributed to most of the P input remaining in farmland soil (Mishima et al., 2003). Applied P is utilized by crops or fixed in complexes. Fixed P may remain in the complex or move to an available pool for crop utilization. Therefore, P for crops comes from P applied in that year or in previous years. Nishio (2003) suggested that the aim in Japan needs to be reductions in P fertilizer input because of the enrichment of soil-available P and the higher levels of P applied than in other countries. Therefore, there is great interest in the solubilization of bound phosphates in soil that have accumulated because of the excessive and repeated use of chemical P fertilizers.
Increases in the levels of P released from phosphate minerals by microbial inoculants are an important aspect of agriculture. The ability of PSM to solubilize bound phosphates and increase the availability of P for rice is a possible mechanism for plant growth promotion under field conditions (Verma et al., 2001). Previous studies reported that the application of PSM, either alone or in combination, to paddy fields enhanced growth and reduced the application of chemical fertilizers. The inoculation of rice grown in Iranian paddy fields with single phosphate solubilizing bacteria (PSB) enhanced rice growth and reduced the application of triple super phosphate by 67% (Bakhshandeh et al., 2015). Furthermore, PSM in combination with other plant growthpromoting rhizobacteria improved agronomic traits, such as root lengths and the biomass of rice grown in organic paddy fields across India (Sherpa et al., 2021). Therefore, PSM may be used as microbial inoculants for rice production.
The present study investigated the potential of inorganic phosphate solubilizing bacteria (iPSB) dwelling in the soil around rice roots for development as microbial inoculants. Soils from paddy rice fields were collected across different regions of Japan. A sequential P extraction analysis was then performed on soil samples. iPSB were isolated using different inorganic P sources. Previous studies on the isolation of PSM only used tricalcium phosphate. In the present study, aluminum phosphate and iron phosphate were also used to isolate iPSB candidates for inoculant development. This approach allowed for the isolation of iPSB and their application to acidic or alkaline soils. To the best of our knowledge, this is the first study to survey iPSB from paddy rice fields across different regions of Japan. The isolates obtained were genetically characterized and their phosphate solubilizing and plant growth-promoting abilities were evaluated.

Soil sampling and sequential P extraction
Soils adhering to roots (SAR) were sampled from paddy fields at eight locations in Fukushima, Hokkaido, Honmachi, Kagawa, Nagano, and Saga, Japan ( Fig. 1 and Table S1). Soils were sampled at depths of 0-20 cm and soil pH was assessed using a 1:5 ratio of soil to deionized water (van Reeuwijk, 2002). The depth of sampling considered the active part for the fibrous roots of rice plants. SAR were collected and stored at 4°C until needed.
Air-dried soils were subjected to HNO 3 -HClO 4 digestion and TP was quantified using the molybdate blue method (Murphy and Riley, 1962). Soil P was sequentially fractioned using a modified Hedley fractionation protocol (Hedley et al., 1982;Imai et al., 2019). Soil (0.5 g) passed through a 2-mm soil sieve was added to a 50-mL centrifuge tube and sequentially extracted by adding 30 mL of each extractant solution. Extractant solutions were added in the following order: distilled water with two resin strips (anion exchange resins in the bicarbonate form), 0.5 M NaHCO 3 (pH 8.5), 0.1 M NaOH, and 1 M HCl. In all extractions, tubes were shaken for 16 h at 25°C. After each extraction, tubes were centrifuged at 1,339×g for 20 min. Supernatants were filtered (5B Advantec) and P concentrations were measured using the molybdate blue method after pH adjustments of the solution. Absorbable and loosely bound P was extracted using a resin strip and 0.5 M NaHCO 3 , Alor Fe-bound P was extracted using 0.1 M NaOH, and Ca-bound P was extracted using 1 M HCl. TP in the NaHCO 3 and NaOH fractions was digested using (NH 4 ) 2 S 2 O 8 in an autoclave at 120°C for 60 and 90 min, respectively. P concentrations were measured using the molybdate blue method. The difference between TP and the sum of all fractionated P was defined as residual P. Damo et al.

Estimation of phosphate solubilizing ability
A total of 147 bacterial isolates were grown in TS broth at 28°C for 24 h. Bacterial cells were centrifuged at 10,000×g for 15 min. The supernatant was discarded and cells were resuspended in 0.85% NaCl solution at a cell density of 10 7 colony forming units mL -1 (CFU mL -1 ). Resuspended cells (10 μL) were transferred to medium plates containing NBRIP supplemented with 5 g L -1 Ca 3 (PO4) 2 for Ca-P isolates, Reyes medium with AlPO 4 for Al-P isolates, or Reyes medium amended with FePO 4 for Fe-P isolates. There were three replicates of each isolate type. Plates were incubated at 30°C for 7 days. The formation of a halo zone around the bacterial colony indicated that the isolate solubilized phosphate. The phosphate solubilizing ability of each isolate was evaluated by measuring the size of the halo zone, and the phosphate solubilization (PS) index was calculated as previously described (Premono Edi et al., 1996). Isolates that showed high PS indices were selected for genetic characterization and plant assays.
A second screening was performed to quantitatively estimate the phosphate solubilizing ability of isolates. Representative isolates with high PS indices were grown in TS broth at 28°C for 24 h and cell density was adjusted to 10 8 CFU mL -1 . The bacterial suspension was then transferred to P growth media supplemented with Ca 3 (PO4) 2 , AlPO 4 , or FePO 4 . After a 5-day incubation at 28°C under shaking at 130 rpm, the supernatant was collected by centrifugation at 13,000×g for 10 min. Soluble P released in the supernatant was quantified using the molybdenum blue assay (Murphy and Riley, 1962). The experiment was performed in triplicate. Uninoculated samples served as negative controls. The positive control bacterium used was Priestia (Bacillus) megaterium NBRC 15308, which is a known phosphate solubilizer (Rodrıǵuez and Fraga, 1999), and was supplied by the National Institute of Technology and Evaluation (NITE) Biological Resource Center.

Extraction of genomic DNA
Isolates were grown in TS broth at 28°C for 24 h. Cells were collected by centrifugation at 13,000×g for 2 min and washed with sterile distilled water. Genomic DNA was extracted using a Wizard ® Genomic DNA Purification Kit (Promega) according to the manufacturer's protocol.

DNA amplification and sequencing
The 16S rRNA and rpoB regions of 63 out of the 147 isolates were sequenced. A polymerase chain reaction (PCR) and sequencing of the 16S rRNA gene were performed using the bacterial universal primers 1F (5′-AGT TTG ATC CTG GCT C-3′) and 3R (5′-AAG GAG GTG ATC CAG CC-3′), (Habibi et al., 2014), and rpoB-F (5′-ATCGAAACGCCTGAAGGTCCAAACAT-3′) and rpoB-R (5′-ACACCCTTGTTACCGTGACGACC-3′) for the rpoB gene (Mohkam et al., 2016). Genes were amplified using 10 ng of purified DNA and KOD Plus Neo (Toyobo). PCR products were purified using a FastGene™ Gel/PCR Extraction Kit (Nippon Genetics) and then sequenced using an ABI PRISM 3500 Genetic Analyzer (Applied Biosystems) according to the manufacturer's instructions. The sequences obtained were compared to the 16S rRNA and rpoB gene sequences deposited in the GenBank database using BLAST online software (http://www.ncbi.nlm.nih.gov/ BLAST). Phylogenetic trees based on the nucleotide sequences of the 16S rRNA and rpoB gene sequences were aligned using Genetyx version 11 (Genetics). Phylogenetic trees were constructed based on the neighbor-joining algorithm with 1,000 replications using the bootstrap method and the maximum composite likelihood model without topology. These processes were conducted using Molecular Evolutionary Genetics Analysis (MEGA) software (version 6.0; Pennsylvania State University, State College, PA, USA) (Tamura et al., 2013).

Nucleotide sequence accession numbers
DNA sequences were deposited in the DNA Data Bank of Japan (DDBJ) under accession numbers LC661699 to LC661761 (16S rRNA) and LC661762 to LC661819 (the rpoB gene).

Plant growth assay
Sixty-three representative isolates were cultured in TS broth at 28°C for 24 h with constant shaking and bacterial cells were then collected by centrifugation and resuspended in 0.85% NaCl solution. Oryza sativa L. 'Koshihikari' (rice) seeds were surfacesterilized by soaking in 70% ethanol for 30 s followed by 3% sodium hypochlorite for 3 min. They were then washed five times with sterile distilled water. The sole insoluble P source of 85 mg of Ca 3 (PO 4 ) 2 for Ca-P isolates, AlPO 4 for Al-P isolates, and FePO 4 for Fe-P isolates was amended in 1 kg of sterilized vermiculite (Subero et al., 2016). Sterilized seeds were incubated at 28°C in the dark for 48 h before germination. Pre-germinated seeds were transplanted into tubes (outer diameter: 29.1 mm×length: 114.4 mm) with 20 g vermiculite (Vermitech), and seeds were then inoculated with 200 μL of the bacterial culture at 10 8 CFU mL -1 . Plants were grown in a growth chamber room under controlled environmental conditions with 16 h light (5,000-7,000 lx) and 8 h darkness, and were maintained at 28±2°C. Each container was irrigated with sterile distilled water at a 100% water-holding capacity for vermiculite throughout the cultivation period. The experiment was a completely randomized design with three replicates for each treatment, and each treatment contained six plants. Uninoculated plants and plants inoculated with Pseudomonas veronii JR37 were used as negative and positive controls, respectively (Habibi et al., 2014). Plants were harvested after 2 weeks and the roots were washed with tap water to remove adhering vermiculite. Fresh shoot and root lengths were measured. Shoots and roots were then dried at 80°C for 48 h before measuring their dry weights. Comparisons between the control and treatment groups were performed using a one-way ANOVA and Dunnett's post hoc analysis at P<0.05 by SAS software (SAS Institute).

Soil P analysis
Sequential P extraction was performed on eight rice soil samples. The TP of soil samples ranged between 6,076 and 21,638 mg P kg -1 . Table S2 shows that SAR from Honmachi had the highest TP among the soil samples examined, while SAR from Fukushima site 2 had the lowest.
The percentages of different P fraction contents in each fraction to TP are shown in Fig. 2. Absorbable P accounted for only 1.6% of TP on average, and both sites Phosphate Solubilizing Strains from Rice in Fukushima had the highest absorbable P percentages (2.6%), whereas Nagano (0.7%) had the lowest. The two SAR samples from Fukushima had significantly different absorbable P values from the average for all soil samples. In contrast, SAR from Nagano and Honmachi had significantly lower absorbable P values. Al-or Fe-bound P ranged between 295 and 750 mg P kg -1 , corresponding to between 7.0 and 2.8% TP. Both SAR from Fukushima had significantly higher percentages of Al-Fe bound P than the average for all soil samples. In addition, Ca-bound P at Fukushima site 2 had the highest Ca-bound P percentage to TP (3.2%), while Kagawa site 1 had the lowest (1.0%). Residual P comprised 92% TP averaged over all soil samples, ranging between 87.3% (Fukushima site 1) and 94.7% (Nagano). In contrast to absorbable P, both samples from Fukushima had significantly lower residual P values than the average for all samples.

Isolation of iPSB
Eight paddy rice SAR were used to isolate the three types of iPSB. The number of phenotypically different isolates obtained per SAR was noted; 147 isolates were obtained (Table 1). Sixty isolates solubilized tricalcium phosphate (Ca-iPSB), 30 solubilized aluminum phosphate (Al-iPSB), and 57 solubilized iron phosphate (Fe-iPSB). The largest numbers of iPSB were isolated from Kagawa site 2, which was an Entisol and organically fertilized. Among all isolates, the highest percentages of Ca-iPSB were recorded for Fukushima site 1 and Kagawa site 2. Fukushima site 1 was classified as an Andisol and was managed using chemical fertilizers. The highest ratio of Al-iPSB was recorded at Kagawa site 1, whereas Fe-iPSB were isolated from Nagano. Soil samples from both sites were categorized as Entisols with organic fertilizer management.
The relationship between the P content in each fraction of soil samples and isolated iPSB was analyzed ( Table 2). The Ca-bound P content negatively correlated (r=-0.74) with the number of isolated Al-iPSB. A negative correlation (r=-0.75) was also observed between the number of isolated Ca-iPSB and Fe-iPSB. In addition, a positive correlation (r=0.91) was noted between absorbable P and Al-Fe-bound P, while a negative correlation (r=-0.91) was found with the residual soil P content. Negative correlations were also observed between the residual P content and Al-Fe bound P (r=-0.99) and Ca-bound P (r=-0.86). Al-Fe-bound P positively correlated with Ca-bound P (r=0.81).

Estimation of phosphate solubilizing ability
PS indices were measured for 147 iPSB. Thirty out of 60 Ca-iPSB with a high phosphate solubilizing ability (Fig. S1A and Table 1) were selected for genetic characterization and plant assays. JC1 isolated from Fukushima site 1 had the highest PS index (2.90) among the representative isolates, whereas JC24 (1.28) from Saga had the lowest. Furthermore, 16 representative isolates with a high phosphate solubilizing ability were selected from 30 Al-iPSB samples ( Fig. S1B and Table 1). JA6 (6.98) from Kagawa site 1 had the highest PS index out of the Al-iPSB isolates. Seventeen  S1C and Table 1).
To further confirm the phosphate solubilizing ability of the representative isolates, the quantification of released P from insoluble P sources was performed. Seventeen isolates exhibited significantly higher solubilized P from Ca 3 (PO 4 ) 2 than the positive control (Fig. 3A). JC2 from Fukushima site 1 had the maximum PS (400 μg mL -1 ) among the representative isolates. In addition, 10 isolates exhibited the significant release of soluble P from AlPO 4 over NBRC 15308 (Fig. 3B). JA9 obtained from Kagawa site 1 had the highest amount of soluble P (76 μg mL -1 ) out of the Al-iPSB isolates. Five isolates had a larger amount of P solubilized from FePO 4 than the positive control (Fig. 3C). Isolate JF3 from Fukushima site 1 had the highest amount of soluble P (53 μg mL -1 ) for the Fe-iPSB group.

Genetic characterization based on 16S rRNA and rpoB sequencing
Representative isolates were subjected to 16S rRNA and rpoB sequence analyses. Sixty-three isolates were selected based on the SAR source and their phosphate solubilizing ability (Table 1 and Fig. S1). The closest reference strains (with >98% similarity) for each isolate were summarized in the phylogenetic analysis (Fig. 4) and their phenotypes with PS indices are shown in Table 3.
One housekeeping gene (rpoB) was phylogenetically analyzed to confirm the identity of iPSB (Fig. S3). In general, the results obtained were consistent with 16S rRNA results. Most of the isolated iPSB were β-Proteobacteria. However, the rpoB gene was not successfully amplified from Actinobacteria or the genus Pleomorphomonas. To the best of our knowledge, the genera Pleomorphomonas (JA1 and JA2), Rhodanobacter (JF8), and Trinickia (JC30, JF12, and JF16) were described herein for the first time as phosphate solubilizing bacteria.

Soil characteristics and iPSB genera
Comparisons of isolated iPSB genera were based on the type of inorganic P source used in isolation, fertilizer management, and the soil type of SAR sources (Fig. 5). Acidovorax, Enterobacter, and Pantoea were obtained after Ca-P isolation, whereas Pleomorphomonas, Ralstonia, Rhodococcus, and Sphingomonas were uniquely isolated using Al-P (Fig. 5A). In addition, Fe-P isolation yielded Cupriavidus, Methylobacterium, Mycolicibacterium, Rhodanobacter, and Variovorax. A few bacterial genera were also shared by two types of inorganic P sources, such as Trinickia from Ca-P and Fe-P isolation, Pseudomonas and Priestia from Ca-P and Al-P isolation, and Paraburkholderia from Al-P and Fe-P isolation. Burkholderia was isolated from inorganic P sources.
The bacterial genera Burkholderia, Cupriavidus, Enterobacter, Pantoea, Paraburkholderia, Priestia, Pseudomonas, and Trinickia were isolated regardless of the fertilizer management applied to the SAR source (Fig. 5B). In contrast, Acidovorax, Methylobacterium, Pleomorphomonas, and Mycolicibacterium were only isolated in paddy fields in which chemical fertilizers had been used, whereas Ralstonia, Rhodanobacter, Rhodococcus, Sphingomonas, and Variovorax were exclusively obtained from paddy fields managed with organic fertilizer.
The soil type of the Rh source was also considered to clarify whether it had any effects on the type of iPSB isolated (Fig. 5C). Rhodococcus and Variovorax were isolated from Inceptisols, whereas Acidovorax, Methylobacterium, Mycolicibacterium, and Pleomorphomonas were found in Andisols. Burkholderia, Ralstonia, Rhodanobacter, Sphingomonas, and Trinickia were isolated from Entisols, and Cupriavidus, Enterobacter, Pantoea, and Paraburkholderia were found in Andisol and Entisol soils. Pseudomonas and Priestia were obtained from all soil types.

Effects of bacterial inoculation on rice plant growth promotion
To evaluate the effects of iPSB on rice plant growth, representative isolates (63) were inoculated onto O. sativa Koshihikari planted in sterilized vermiculite amended with insoluble P. P. veronii JR37 (AB793766; Habibi et al., 2014) was also included in the plant assay as a positive control.
A summary of rice plant growth promotion (PGP) by iPSB is shown in Table 4. Regardless of the insoluble P used, some isolated iPSB performed better than the positive control P. veronii JR37. In the Ca-iPSB plant assay using Ca 3 (PO 4 ) 2 as the sole P source, the maximum dry biomass (shoot, root, and total) was recorded for rice plants inoculated with Acidovorax sp. JC1 (Fig. S4). In comparisons with plants that had not been inoculated with Ca-P (UN+Ca-P), root lengths were significantly longer in rice plants inoculated with Acidovorax sp. JC5 and Pseudomonas sp. JC11. These isolates were isolated from Fukushima: Acidovorax from Site 1 and Pseudomonas from Site 2.
The results of the AlPO 4 assay showed that rice plants inoculated with Burkholderia sp. JA6 from Kagawa site 1 and Sphingomonas sp. JA11 from Nagano had the longest root lengths (Fig. S5). In addition, rice plants inoculated with the Kagawa site 2 isolate Burkholderia sp. JA10 has significantly longer root lengths than those that had not been inoculated with Al-P (UN+Al-P). Among the Al-iPSB used, Burkholderia sp. JA9 from Kagawa site 1 produced the highest shoot and total dry weights. Rice plants inoculated with Ralstonia pickettii JA12 from Nagano had the highest root dry weight.
The effects of Fe-iPSB on rice were examined using FePO 4 as an insoluble P source. Shoot lengths were affected by six isolates from 17 representative Fe-iPSBs (Fig. S6). Mycolicibacterium sp. JF3 had the longest shoot length among Fe-iPSB. Moreover, 16 isolates had longer root lengths than the insoluble P-amended control. Rice plants inoculated with Variovorax sp. JF6 (Hokkaido) and Mycolicibacterium sp. JF5 (Honmachi) had significantly longer root lengths than UN+Fe-P. Rice plants inoculated with Rhodanobacter sp. JF8 from Nagano had the highest shoot dry weights, whereas Burkholderia sp. JF14 from Kagawa site 1 had the highest root dry weight. Rice plants inoculated with Burkholderia sp. JF15 (Kagawa site 2) had the highest total dry weight.
The majority of promising iPSB were isolated from paddy fields with organic fertilizer management (Table 4). Additionally, the top isolates showed longer root lengths of the rice cultivar Koshihikari with insoluble P as the sole P source than the control.

Soil P content and iPSB
The soil P content and its forms were evaluated in the SAR samples used to isolate the three sets of iPSB ( Fig.  2 and Table S2). P bioavailability and stability were previously shown to significantly depend on soil characteristics (Wright, 2009). Among the soil samples examined, SAR from Honmachi had the highest TP content. It was an Andisol, which generally has a high P fixation capacity. However, SAR from Fukushima site 1 was also an Andisol, but had a low TP content. SAR from Saga, classified as an Entisol, had the lowest TP content. Entisols have inherently low TP and a poor P fixation capacity (Dissanayaka et al., 2015). Furthermore, the P forms that had accumulated in croplands in which P was in predominantly inorganic forms were affected by the application of fertilizers and manure (Ando et al., 2021). Ca-bound P at Fukushima site 2 was present at the highest percentage among TP in all samples. In this field, chemical fertilizers were applied during the

Fig. 5. Venn diagram for isolated iPSB genera
Unique and/or shared genera among the types of inorganic P used to isolate bacteria (A), between fertilizer management regimes (B), and soil types (C) in paddy field sources. Al-P: AlPO 4 , Ca-P: Ca 3 (PO 4 ) 2 , Fe-P: FePO 4 rice-planting season. The lowest percentage of Ca-bound P was observed at Kagawa site 1, which was an organically cultivated site. This is consistent with the findings by Wright (2009) who reported that the application of superphosphate (a chemical fertilizer) increased the Ca-P content in the surface soils of agricultural land. The present results also revealed a positive correlation between the absorbable P content with Al-P and Fe-P in soil samples, as detected in SAR from Fukushima, which had the largest absorbable P, Al-P, and Fe-P contents. This is in agreement with previous findings showing that labile and relatively stable P forms were influenced by Al and Fe in Japanese soils (Tani et al., 2010;Nakamura et al., 2019).
iPSB were isolated from paddy rice SAR and may be used to develop microbial inoculants. A selective medium without available soluble P sources was used to isolate potential iPSB. Soil pH affects the P forms present, which are mainly Ca-bound P in alkaline soils and Al or Fe complexes in acidic soils. Therefore, three insoluble P sources (Ca 3 [PO 4 ] 2, AlPO 4 , and FePO 4 ) were used to isolate iPSB, yielding three sets of iPSB. Interestingly, there was no universal metal-P compound for the selection of PSM (Bashan et al., 2013). Previous studies reported that the isolation of iPSB using Ca 3 (PO 4 ) 2 alone yielded iPSB with a low to no solubilizing ability with AlPO 4 and FePO 4 (Pérez et al., 2007;Park et al., 2010). These findings imply that isolated iPSB are impractical for acidic soils.
Environmental factors generally influence the distribution of bacterial communities. Therefore, the relationship between the soil P content and number of isolated iPSBs was examined. Despite the relatively homogenous soil pH, different types of iPSB, such as those associated with inorganic P (Ca-P, Al-P, and Fe-P), were isolated. Therefore, the types and number of iPSB may be affected by the fertilizer management regime and the soil type of paddy field sources. Among all isolates, the majority of iPSB were obtained from Entisol soils. Although this soil has a low Al/Fe oxide content (Dissanayaka et al., 2015), a large portion of Al-iPSB and Fe-iPSB were isolated from SAR in this soil type. With the exception of Fukushima site 1, a high percentage of Ca-iPSB was taken from Kagawa site 2 with low Ca-bound P. Specific iPSB are more likely to be isolated based on the low content of inorganic P content in soil, such as the high percentage of Al-iPSB originating from soil with low Al-P and Fe-P contents. The discrepancy of high Ca-iPSB in high Ca-bound P soil in Fukushima may be explained by the chemically fertilized nature of the field. The fertilizer management of paddy rice SAR samples must have affected the occurrence of iPSB. Previous studies demonstrated that the application of chemical fertilizers may reduce the abundance of beneficial taxa for plant health and growth (Sinong et al., 2021). This was reflected in the present results, with the highest portion of total iPSB being obtained at Kagawa sites 1 and 2 and at Nagano, at which organic farming is practiced.

Characterization of iPSB
A total of 147 iPSBs were screened based on their PS indices. Sixty-three isolates were selected in consideration of the SAR from which they were obtained. The results of 16S rRNA and rpoB gene sequencing led to iPSB being classified into five groups (Fig. 4, S2 and S3).
The first group was Actinobacteria, which was composed of Rhodococcus and Mycolicibacterium. Rhodococcus sp. JA4 from Hokkaido solubilized AlPO 4 , whereas the isolates Mycolicibacterium sp. JF3 and JF5 solubilized FePO 4 . The genus Rhodococcus possesses plant growth-promoting properties and is commonly isolated in soils polluted with heavy  (Belimov et al., 2001) and petroleum hydrocarbons (Pacwa-Płociniczak et al., 2016). Although R. erythropolis has been shown to solubilize Ca 3 (PO 4 ) 2 , its ability to solubilize AlPO 4 has not yet been demonstrated. The phytoremediation abilities shown by Rhodococcus have been examined (Toussaint et al., 2012;Álvarez-López et al., 2016). Since this genus degrades persistent pollutants, isolated Rhodococcus exhibits tolerance to heavy metals, such as aluminum. Mycolicibacterium may also be obtained from soil contaminated with polycyclic aromatic hydrocarbons (Hormisch et al., 2004). Although M. fluoranthenivorans has been applied to food and feed processing, specifically to the degradation of aflatoxin B1 (Hormisch et al., 2004;Teniola et al., 2005), its potential as a PGP bacterium was described herein for the first time.
Priestia (formerly known as Bacillus) belongs to the Bacilli group. The majority of isolated Priestia species were Ca-iPSB; however, there was one Al-iPSB. This genus is considered to be a strong solubilizer of tricalcium phosphate (Rodrıǵuez and Fraga, 1999;Li et al., 2019). Previous studies showed that members of the genus exhibited the ability to solubilize P from aluminum phosphate (Banik and Dey, 1983). Methylobacterium, Pleomorphomonas, and Sphingomonas are known plant growth-promoting bacteria and comprise the α-Proteobacteria group of isolated iPSB. The plant growth-promoting mechanism includes the production of 1-aminocyclopropane 1-carboxylate deaminase in M. fujisawaense and M. oryzae (Madhaiyan et al., 2006(Madhaiyan et al., , 2007; however, the phosphate solubilizing ability of M. fujisawaense has not yet been reported. The genus Pleomorphomonas was herein identified for the first time as a phosphate solubilizing bacteria. This genus is generally associated with nitrogen-fixing bacteria, such as Pleomorphomonas oryzae, which has also been isolated from paddy rice fields (Xie and Yokota, 2005).
The γ-Proteobacteria group consisted of Pseudomonas, Enterobacter, Pantoea, and Rhodanobacter. Some of the isolated Pseudomonas had the capacity to solubilize P from aluminum phosphate. Isolated Enterobacter and Pantoea were exclusively obtained using tricalcium phosphate, a common inorganic P source for the isolation of iPSB. These genera are typically referred to as phosphate solubilizers. However, it is important to note that the present results showed that Rhodanobacter functioned as an iron phosphate solubilizer. The type strain of this bacterium was isolated from a contaminated aquifer with a high nitrate content and was capable of complete denitrification (Prakash et al., 2012).
The majority of iPSB were classified under β-Proteobacteria as Acidovorax, Burkholderia, Cupriavidus, Paraburkholderia, Ralstonia, Trinickia, and Variovorax. Most of these genera are common phosphate solubilizers, except for Fe-iPSB, which were similar to Variovorax ginsengisoli, reported herein for the first time. V. ginsengisoli is generally considered to be a denitrifying bacterium that is isolated from ginseng (Im et al., 2010). In addition, the genus Trinickia was shown to function as a phosphate solubilizer in this study. This genus is characterized by plant growth-promoting (Fu et al., 2019) and heavy metal-tolerant bacteria (Zhu et al., 2012).

Plant growth promotion of iPSB
Promising iPSB demonstrated their potential for plant growth promotion in rice with insoluble P as the sole P source. In consideration of the phosphate solubilizing ability of these isolates, positive correlations were observed with PS indices, P solubilization, and plant growth promotion in rice. A positive correlation was also noted between PS indices and plant growth traits in Ca-iPSB, specifically shoot dry weight (r=0.38) and total dry weight (r=0.41). Moreover, positive correlations were found between root length and the PS index (r=0.57) and also P solubilization (r=0.80) in Al-iPSB. Therefore, isolation and screening methods based on specific inorganic phosphates were reliable for Ca-iPSB and Al-iPSB as candidate microbial inoculants. However, no significant relationships were observed for Fe-iPSB. This may be attributed to the isolation method, which heavily depended on the production of acids by bacteria. Another mechanism for the liberation of iron in FePO 4 may be through siderophores. Therefore, further studies are needed to confirm the relationship between the PGP and phosphate solubilizing abilities of isolates.
The isolates Acidovorax sp. JC5 and Pseudomonas sp. JC11 increased the root lengths of rice treated with Ca 3 (PO 4 ) 2 and showed the high release of soluble P in vitro. Microbial inoculants may be developed using these isolates for soils that are heavily fertilized with chemical P fertilizers because the dominant P species in these soils may be Ca-P. Although available P is high and sufficient for crops in these paddy soils, a portion of the soluble chemical P fertilizer applied will still be rendered insoluble or leach into bodies of water, potentially causing eutrophication. The aim of microbial inoculants with Ca-iPSB under these conditions is to efficiently utilize and ultimately reduce the amount of the chemical P fertilizer used. Moreover, existing bound P from previous cropping seasons may be utilized again by plants through solubilization. The Al-iPSB isolates Burkholderia sp. JA6 and JA10 and Sphingomonas sp. JA11 improved root lengths and a quantitative estimation revealed higher soluble P than a known P solubilizer by utilizing AlPO 4 . These isolates may be used to develop microbial inoculants for volcanic soils (Andisol), which have a high P fixation capacity due to Al oxides. Root lengths were increased by Fe-iPSB Variovorax sp. JF6 and Mycolicibacterium sp. JF5 with FePO 4 as the only P source. The abovementioned Fe-iPSB may potentially be developed as inoculants for red-yellow soils (Inceptisol or Ultisol) with a strong P fixation ability because they contain Fe oxide. However, further studies on field conditions are needed to confirm the potential of these isolates as microbial inoculants for high P fixation areas.